
ADL5310
Rev. A | Page 15 of 20
A 10 nF capacitor on each VSUM pin (20 nF parallel equivalent)
combined with the 16 kΩ source resistance yields a 500 Hz pole,
which is sufficiently below the bandwidth for the minimum
input current of 3 nA.
Residual crosstalk disturbance is particularly problematic at the
lowest currents for two reasons. First, the loop is unable to reject
summing node disturbances beyond the limited bandwidth.
Second, the settling response at the lowest currents to any
residual disturbance is significantly slower than that for input
currents even one or two decades higher (see
Figure 18).–6
–3
0
3
6
9
12
INACTIV
E
CHANNE
L
OUTP
UT
(mV
)
0
0.2
0.4
0.6
0.8
1.0
1.2
ACTIV
E
CHANNE
L
OUTP
UT
(V
)
0
0.5
1.0
1.5
2.0
2.5
TIME (ms)
04415-0-036
ACTIVE CHANNEL OUTPUT PULSE, 1-DECADE STEP
3
A TO 30A
INACTIVE CHANNEL RESPONSE
IINP – 100nA
IINP – 10nA
IINP – 30nA
IINP – 3nA
Figure 36. Crosstalk Pulse Response for Various Input Current Values
(dc input) to a 1-decade current step on the input of the active
channel for several inactive channel dc current values. Addi-
tional system considerations may be necessary to ensure
adequate settling time following a known transient when one or
both channels are operating at very low input currents.
RELATIVE AND ABSOLUTE POWER
MEASUREMENTS
When properly calibrated, the ADL5310 provides two inde-
pendent channels capable of accurate absolute optical power
measurements. Often, it is desirable to measure the relative
gain or absorbance across an optical network element, such as
an optical amplifier or variable attenuator. If each channel has
identical logarithmic slopes and intercepts, this can easily be
done by differencing the output signals of each channel. In
reality, channel mismatch can result in significant errors over a
wide range of input levels if left uncompensated. Postprocessing
of the signal can be used to account for individual channel
characteristics. This requires a simple calculation of the
expected input level for a measured log voltage, followed by
differencing of the two signal levels in the digital domain for a
relative gain or absorbance measurement. A more straight-
forward analog implementation includes the use of a current
mirror, as shown in
Figure 37. The current mirror is used to
feed an opposite polarity replica of the cathode photocurrent of
PD2 into Channel 2 of the ADL5310. This allows one channel to
be used as an absolute power meter for the optical signal
incident on PD2, while the opposite channel is used to directly
compute the log ratio of the two input signals.
5V
IPD2
IIN2=IPD2
IIN1
TEMPERATURE
COMPENSATION
BIAS
GENERATOR
1k
2M
4.7nF
1k
4.7nF
0.1
F
1k
4.7nF
1nF
0.1
F
1nF
VNEG
COMM
VREF
VRDZ
VPOS
VSUM
INP1
PD1
InGaAs PIN
1k
4.7nF
PD2
InGaAs PIN
IRF1
ILOG1
OUT1
SCL1
BIN1
LOG1
04415-
0-
037
TEMPERATURE
COMPENSATION
COMM
log
5V
VSUM
5V
INP2
IRF2
ILOG2
OUT2
Φ
2*
α
21**
*
Φ
2(V) 0.2log10()
SCL2
BIN2
LOG2
IIN2
100pA
**
α
21(V) 0.2log10()
IIN1
IPD2
ADL5310
Figure 37. Absolute and Relative Power Measurement Application
Using Modified Wilson Current Mirror
The presented current mirror is a modified Wilson mirror.
Other current mirror implementations would also work, though
the modified Wilson mirror provides fairly constant perfor-
mance over temperature. It is essential to use matched pair
transistors when designing the current mirror to minimize the
effects of temperature gradients and beta mismatch.